#7373: [with spkg; needs review] Disable assembly code in libgcrypt on risky
platforms.
------------------------+---------------------------------------------------
Reporter: drkirkby | Owner: drkirkby
Type: defect | Status: needs_review
Priority: major | Milestone: sage-4.2.1
Component: porting | Keywords:
Work_issues: | Author:
Reviewer: | Merged:
------------------------+---------------------------------------------------
Changes (by drkirkby):
* status: new => needs_review
Old description:
> On my Sun Ultra 27, which has a quad core Xeon, running OpenSolaris
> 06/2009, libgrcypt would not build. The error message indicated it was
> related to the use of assembly code.
>
> However, I believe libgcrypt did not cause an issue on 'disk.math', so
> I'm somewhat surprised it did on my Ultra 27. But I think it is safer to
> disable assembly language on all Solaris x86 systems. (It is '''not'''
> necessary to do so on Solaris on SPARC)
>
> I also added some tests for other platforms (AIX, HP-UX, Tru64 and IRIX)
> and disabled assembly language on them too. It is most unlikely assembly
> code for them will work, and I hope to try at least some of these
> platforms in the near future.
>
> The only updates are to spkg-install and SPKG.txt. The revised files will
> be put into
>
> http://sage.math.washington.edu/home/kirkby/Solaris-
> fixes/libgcrypt-1.4.4.p1
>
> within 30 minutes of this post. (I thought I'd get the trac ticket in
> first, so the trac number can go into the SPKG.txt)
>
>
> Dave
New description:
Libgcrypt is a package which has some assembly code routines. An option to
the ''configure'' script
{{{
--disable-asm
}}}
will disable the use of assembly code. This is mandatory on some platforms
such as OS X and 64-bit SPARC, and would appear to be desirable on others,
unless we can be sure it works.
On my Sun Ultra 27, which has a quad core Xeon, running OpenSolaris
06/2009, libgrcypt would not build. The error message indicated it was
related to the use of assembly code.
However, I believe libgcrypt did not cause an issue on 'disk.math', so I'm
somewhat surprised it did on my Ultra 27. But I think it is safer to
disable assembly code on all Solaris x86 systems.
I disabled the assembly code on 64-bit SPARC too, as I know that is broken
- see #7127. Disabling the assembly code does not allow libgcrypt to be
built as 64-bit code on Solaris, but the build does get further with the
assembly code disabled. (It ultimately fails, as both 32-bit and 64-bit
objects are created).
I also added some tests for other platforms (AIX, HP-UX, Tru64 and IRIX)
and disabled assembly language on them all except HP-UX, where I could
verify it did not cause any problems.
'''Hence assembler code is disabled on'''
* OS X (both 32 and 64-bit)
* Solaris x86 (both 32 and 64-bit)
* Solaris SPARC (32-bit only)
* IRIX (as a precaution, as untested)
* AIX (as a precaution, as untested)
* Tru64 (as a precaution, as untested)
'''
Assembly code is enabled on:'''
* Linux
* Solaris SPARC (32-bit only)
* HP-UX
* Anything else, not mentioned above. This would include Cygwin.
'''The updated package has been tested on'''
* sage.math
* bsd.math
* Sun Ultra 27, Intel Xeon processor, running OpenSolaris (x86)
* Sun Blade 2000, SPARC processor in 32-bit mode.
* Sun Blade 2000, SPARC processor in 64-bit mode. The code does not build
fully in this case, as explained above.
* HP C3600 running HP-UX 11.11. The code builds fully.
It should be noted, the HP C3600 uses a PA-RISC processor. More modern HP-
UX machines use the Itanium processor. These would certainly not use the
same assembly code.
The code may be found here.
http://sage.math.washington.edu/home/kirkby/Solaris-
fixes/libgcrypt-1.4.4.p1
Dave
--
--
Ticket URL: <http://trac.sagemath.org/sage_trac/ticket/7373#comment:1>
Sage <http://www.sagemath.org>
Sage: Creating a Viable Open Source Alternative to Magma, Maple, Mathematica,
and MATLAB
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"sage-trac" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/sage-trac?hl=en
-~----------~----~----~----~------~----~------~--~---
